🙊 案例

1. 做家务

我们用上面做家务的例子来用协程实现一下：

import time
import asyncio
async def boil_water():
    print("A: Boil Water")
    time.sleep(2)
    print("A1: I have prepared water(*)")
    await asyncio.sleep(10)
    print("A2: Water has boiled up")
    time.sleep(3)
    print("A3: Water has been poured into the bottle(*)")
async def wash_clothes():
    print("B: Wash Clothes")
    time.sleep(5)
    print("B1: I have prepared clothes(*)")
    await asyncio.sleep(15)
    print("B2: Clothes have been washed clean")
    time.sleep(5)
    print("B3: Clothes have been hanged up(*)")
async def mop_floor():
    time.sleep(10)
    print("C: I have mop the floor clean(*)")
loop = asyncio.get_event_loop()
tasks = [boil_water(), wash_clothes(), mop_floor()]
start = time.time()
loop.run_until_complete(asyncio.wait(tasks))
loop.close()
end = time.time()
print("cost time: ", end - start)

输出结果：

B: Wash Clothes
B1: I have prepared clothes(*)
A: Boil Water
A1: I have prepared water(*)
C: I have mop the floor clean(*)
A2: Water has boiled up
A3: Water has been poured into the bottle(*)
B2: Clothes have been washed clean
B3: Clothes have been hanged up(*)
cost time:  25.004190921783447

显然这和我们预期的结果一样，但实际上我们调整一下 tasks 里面元素的位置，就会发现出来的结果会有所不同，这是因为程序在 CPU 空闲时会随机看哪个任务可以上，就开启哪个任务。

为了得到最理想的结果的话，可以简单做一下改写：

import time
import asyncio
async def boil_water():
    print("A: Boil Water")
    time.sleep(2)
    print("A1: I have prepared water(*)")
    await mop_floor()
    print("A2: Water has boiled up")
    time.sleep(3)
    print("A3: Water has been poured into the bottle(*)")
    async def wash_clothes():
        print("B: Wash Clothes")
        time.sleep(5)
        print("B1: I have prepared clothes(*)")
        await boil_water()
        print("B2: Clothes have been washed clean")
        time.sleep(5)
        print("B3: Clothes have been hanged up(*)")
        async def mop_floor():
            time.sleep(10)
            print("C: I have mop the floor clean(*)")
            loop = asyncio.get_event_loop()
            tasks = [
                loop.create_task(wash_clothes()),
            ]
            start = time.time()
            loop.run_until_complete(asyncio.wait(tasks))
            loop.close()
            end = time.time()
print("cost time: ", end - start)

这样调整之后，每次运行之后的结果就都和我们的预期一致了，但显然这种人工调整的方式在实际场景中并不适用。